Septic shock appears to result from excessive release of cytokines [e.g., tumor necrosis factor-a (TNF-a), IL-2, etc.] and other pro-inflammatory substances [e.g., nitric oxide (NO)] from cells of the monocyte/macrophage lineage in response to infection or lipopolysaccharide (LPS) administration. The production of these cytokines, and their action, is mediated by signal transduction events that induce protein tyrosine phosphorylation. Theoretically, inhibition of protein tyrosine phosphorylation may be beneficial in sepsis. These compounds would block the potentially high cytokine production that is dependent on tyrosine phosphorylation. These protein kinase inhibitors would block both activation and production of cytokines by bacterial products and the effects of cytokines on target cells. Tyrphostins AG 126 and AG 556 are both protein kinase inhibitors and have been shown to improve outcome in small animal models during both LPS and live bacterial challenge. Further, both AG 126 and AG 556 have been shown to inhibit LPS-induced TNF production from dog peripheral blood mononuclear cells, in vitro. In collaboration with Dr. Novogrodsky and his colleagues, we evaluated AG 126 and AG 556 in our previous canine peritonitis model. In a controlled clinical trial in 100 animals over 6 months, AG 556 but not AG 126 significantly improved survival and prevented multiorgan failure during canine septic shock. Recent analysis of animal experimental data suggests that the effect of anti-inflammatory agents is dependent in part on the underlying infectious burden of the animal. It appears that studies in which controls exhibited high mortality showed improved survival in response to anti-inflammatory therapy. Conversely, studies in which controls exhibited lower mortality suggested that anti-inflammatory agents had no benefit, and possibly some harm. Therefore, it is possible that the reason that human clinical trials in sepsis have shown no benefit is that the anti-inflammatory agents have been given to individuals with varying degrees of illness, and that a subgroup of patients with higher burden of illness might be helped by anti-inflammatory therapy. This study is designed to examine the effect of titrating AG 556 to the severity of illness in canines infected with high- and low-infectious burdens. In our canine model of peritonitis, cohorts of animals with either high or low burdens of E. coli peritonitis clots will be studied. We will compare the efficacy with standard dose 2.5 mg/kg AG 556 to placebo, to titrated dosing 1mg/kg and then 1 or 4 mg/kg depending upon the blood pressure of animals at the 4 h time point. This study is the first study in an animal model to examine whether the utility of anti-inflammatory therapy is dependent upon the burden of infectious agent, and has potential implication for human clinical trials of anti-inflammatory agents in sepsis. We expect to be able to perform this study in our new sedated and ventilated model of bacterial pneumonia sepsis towards the end of the year.